WritingRing Dataset

---

license: cc

Data Format and Structure

The data is organized under ./data/{username}/{action}, where each action corresponds to a different handwriting task:

• Action 0: Letters, hand raised
• Action 1: Letters, hand not raised
• Action 2: Words (connected), hand raised
• Action 3: Words (connected), hand not raised
• Action 4: Words (disconnected), hand raised
• Action 5: Words (disconnected), hand not raised

Each action contains multiple data sets, consisting of data from a ring and a touchpad. Due to the size of the touchpad data, it's split into multiple parts. For example, for dataset with ID 1:

• Ring 0 data: 1_ring_0.bin
• Touchpad data: 1_board_{index}.gz
• Ring 1 data: 1_ring_1.bin (Please ignore this.)

Ring Data Format

The ring data is stored in numpy format as 64-bit floating point values. After reshaping with data_ring0 = data_ring0.reshape(-1, 7), the columns represent:

1. Acceleration X
2. Acceleration Y
3. Acceleration Z
4. Angular velocity X
5. Angular velocity Y
6. Angular velocity Z
7. Timestamp

The data is sampled at 200Hz.

Touchpad Data Format

The touchpad data is stored in gzipped format (.gz).

Data Visualization

Ring Data Visualization

The ring_plot.py script can be used to visualize the ring data. It contains a function plot_ring_data(), which takes two parameters:

1. Path to the ring data file
2. A flag indicating whether to save the output

The output is a plot that visualizes the ring data.

Touchpad Data Visualization

The board_plot.py script is used for visualizing touchpad data. It contains a function plot_board_data(), which takes two parameters:

1. Path to the touchpad data file
2. Username

The function randomly samples touchpad data and outputs a visualization showing the user's writing trajectory, with color representing the pressure applied. Additionally, the visualize_touch_data() function can be called to visualize touchpad data for a specific time range.